Method and system for reallocating data in a file system

Abstract

A method and system for reallocating the data blocks of a logical data entity are described. According to one aspect of the invention, the physical arrangement of data blocks of a logical data entity are analyzed to determine a fragmentation level. Next the fragmentation level is compared with a fragmentation threshold previously assigned to the logical data entity. If the fragmentation level exceeds the fragmentation threshold, only those data blocks of the logical data entity that are within a predefined allocation area having a predetermined number of contiguous data blocks that are not associated with the logical data entity are reallocated.

Description

FIELD OF THE INVENTION[0001]The present invention relates generally to network-based storage systems, and in particular, to a method and system for reallocating data blocks of a logical data container, such as a file, volume or LUN, in an active file system of a network-based storage system with write anywhere capabilities.BACKGROUND[0002]In a typical modern file system, when a file is initially written to a disk or other mass storage device, the file is divided into equally-sized, logical data blocks, and an attempt is made to write the logical data blocks sequentially and contiguously to the disk or mass storage device. Consequently, when the file is read from the disk or mass storage device, the individual data blocks can be read sequentially, thereby limiting the number of read operations required, which in turn limits the seek time required by the reading component (e.g., disk drive head) to search for and locate the individual data blocks comprising the file. However, over tim...

AI Technical Summary

Problems solved by technology

Consequently, when the file is read from the disk or mass storage device, the individual data blocks can be read sequentially, thereby limiting the number of read operations required, which in turn limits the seek time required by the reading component (e.g., disk drive head) to search for and locate the individual data blocks comprising the file.

However, over time, as existing files are modified and deleted, and new files are written, the availability of physical contiguous space for writing new files becomes limited.

Consequently, files tend to become fragmented.

When the files of a file system become fragmented, multiple read operations are required to read the files, and the overall performance of the system suffers.

For a variety of reasons, fragmentation issues may arise with network-based storage systems that implement write anywhere file systems, such as the WAFL® (Write Anywhere File Layout) file system developed by Network Appliance of Sunnyvale, Calif.).

Consequently, with a write anywhere file system, certain classes of workloads are likely to endure performance degradation due to fragmentation issues.

For instance, performance levels associated with workloads that are characterized by a large number of random writes followed by a sequential read tend to suffer from fragmentation issues.

For example, during a work day, a large file that makes up an individual user's email mailbox may be modified several times, causing the large file to become severely fragmented.

However, due to the fragmentation caused by the user's activity throughout the day, the system may have a difficult time sequentially reading the mailbox file.

However, the defragmentation operation does not increase the number of allocated disk blocks.

Traditional defragmentation applications may be ineffective when used with certain network-based storage systems that implement write anywhere file systems.

In particular, the WAFL® file system has certain characteristics, which render traditional defragmentation tools less useful.

However, because data blocks “SN1”, “A”, “B”, “C” and “D” are part of the read-only Snapshot, those data blocks cannot be moved or rewritten.

Consequently, a traditional defragmentation application, which might move the data blocks of the active file system to a new location, would be ineffective because it would nearly double the number of allocated blocks.

Furthermore, in an environment with one or more network-based storage systems, there are often multiple file systems involved.

However, by improving the logical arrangement, the locally executed defragmentation application may cause greater fragmentation of the data blocks as physically stored on the mass storage device.

Consequently a defragmentation application executed locally on the client may actually make matters worse in certain instances.

Images